EI-ERIM-OR seminar May 30

12:00-12:30: Dr. Chris Tzagkarakis (ICS-FORTH)

Intelligent and trustworthy IoT systems: Challenges and solutions in forecasting, explainability, and security 

Abstract: The rapid proliferation of Internet of Things (IoT) devices has led to an unprecedented influx of data, necessitating advanced methods for data analysis, security, and system transparency. This talk explores the challenges and emerging solutions associated with developing intelligent and trustworthy IoT systems. We begin by examining time series forecasting in resource-constrained IoT environments, highlighting techniques for accurate predictions of time-series data. Building on this, we discuss the role of explainable AI (XAI) in enhancing the transparency and trustworthiness of machine learning models within IoT frameworks. Finally, we address critical security issues, focusing on recent approaches for detecting and mitigating botnet attacks. By integrating insights from research and practical applications, this talk aims to provide an overview of the current landscape and future directions in creating secure, efficient, and transparent IoT systems.

12:30-13:00: Stelios Roubakis (ICS-FORTH)

Cloud Native Technologies for Decision Intelligence Systems

Abstract: The increasing complexity of data-rich environments has amplified the demand for systems capable of intelligent, adaptive, and optimized decision-making. Decision Intelligence Systems (DIS) address this need by integrating Artificial Intelligence (AI) and Operations Research (OR), combining data-driven learning with rigorous optimization techniques to support high-impact decisions across diverse domains. This talk examines the role of cloud native technologies in enabling scalable, modular, and automated deployments of DIS. Beginning with the architecture of a modern decision pipeline, the discussion highlights how AI and OR models are developed, orchestrated, and continuously updated within cloud native ecosystems. Special attention is given to the infrastructure patterns that support real-time decision automation, continuous learning, and system resilience.